pdf-to-markdown

```bash

cd .claude/skills/pdf-to-markdown

npm install

```

Dependencies: pdf-parse

```bash

# Basic conversion

node .claude/skills/pdf-to-markdown/scripts/convert.cjs \

--file ./document.pdf

# Custom output path

node .claude/skills/pdf-to-markdown/scripts/convert.cjs \

--file ./doc.pdf \

--output ./output/doc.md

```

| Option | Required | Description |

| ----------------- | -------- | ------------------------------------------------ |

| --file | Yes | Input PDF file |

| --output | No | Output Markdown path (default: input name + .md) |

```json

{

"success": true,

"input": "/path/to/input.pdf",

"output": "/path/to/output.md",

"wordCount": 1523,

"warnings": ["Tables may not be accurately converted"]

}

```

• Text extraction from digital PDFs
• Headings (detected by font size heuristics)
• Paragraphs
• Basic lists
• Links (when embedded in PDF)

• Tables: Very limited support; may not render correctly
• Multi-column layouts: Text may interleave between columns
• Scanned PDFs: NOT supported (requires OCR - see alternatives below)
• Images: NOT extracted (PDF images are not included in output)
• Complex formatting: May be simplified or lost
• Password-protected PDFs: NOT supported

For scanned PDFs (OCR needed):

• Use scribe.js-ocr library (AGPL license)
• Commercial OCR services (Google Cloud Vision, AWS Textract)

For complex tables:

• Consider AI-based extraction (LLM post-processing)
• Manual review and correction

For image extraction:

• Use unpdf library with sharp for image extraction
• Process images separately and reference in markdown

Dependencies not found: Run npm install in skill directory

Empty output: PDF may be scanned/image-based (requires OCR)

Garbled text: PDF may use embedded fonts not supported by parser

Memory issues: Large PDFs may require --max-old-space-size=4096 flag

• Always plan and break many small todo tasks
• Always add a final review todo task to review the works done at the end to find any fix or enhancement needed